“Duplicate, Google chose different canonical than user” is the Search Console status where you told Google which URL should represent a set of duplicate pages, and Google looked at the same pages and picked a different one. The page carrying this status is not indexed; the URL Google chose instead is. Of the three canonical statuses in the Page indexing report, this is the only one that records a disagreement, which is why it deserves a closer look than the other two before you decide whether anything is wrong.
Why Google can overrule your canonical tag
Because the tag was never a directive. Google’s canonicalization documentation calls rel="canonical" a strong hint, weighed alongside other signals: redirects, sitemap inclusion, internal linking, HTTPS over HTTP, and whether the URLs actually serve equivalent content. When the other signals outvote your tag, Google’s choice wins, and this status is where the loss is reported.
That framing tells you what the status actually is: not a syntax error, but a signals disagreement. Your tag said one thing, the rest of your site said another, and Google sided with the rest of your site.
What Google’s own diagnostic says
The Page indexing report documentation gives a four step check for this exact status, and two sub-cases inside it are the ones people miss:
- Inspect the affected URL and read Google-selected canonical.
- Compare it with your User-declared canonical in the same panel.
- Open all three pages in your browser: the tested page, your declared canonical, and Google’s selection.
- Then apply Google’s logic. The status means Google thinks the tested page is a duplicate of the Google-selected URL, not of the URL you declared. Two special readings follow from that. If the Google-selected canonical is the tested page itself, Google thinks the page is not a duplicate of anything, and your tag pointed a unique page at some other URL. And if your declared canonical is not actually similar to the page carrying the tag, Google will never choose it, because a duplicate must be similar to its canonical. That is what duplicate means.
That second sub-case is the one I meet most in audits: a template stamps every variant of a page with a canonical pointing at a parent category or a homepage, hoping to consolidate signals upward. Google does not treat that as consolidation, it treats it as a false claim, ignores the tag, and files the disagreement under this status.
The shapes this status takes on real sites
Almost every case I have seen falls into one of four:
Google picked a protocol or host variant of the same page. Your tag says the non-www URL, but internal links, the sitemap and your redirects all say www, so Google selects www. This one is a self-inflicted signals conflict: fix the tag to match everything else, not the other way round. It is the same consistency argument I walked through when auditing my own site’s trailing slashes: tag, internal links and sitemap should name the same form.
The declared canonical points somewhere dissimilar. The stamped-template case above. Google’s fix is blunt: either make the pages genuinely similar or stop declaring the false canonical.
Cross-domain syndication. Your article runs on a partner site, both carry your canonical, but the partner version has more links and Google picks it. The tag is a hint here too, and a well-linked republication can outvote the original.
Near-identical regional or parameter variants. Storefronts and category filters where the pages really are duplicates and Google consolidates them differently than you hoped. On a country storefront this is usually not a tag problem at all, which is the point of hreflang errors and why country versions are not indexed. Whether that is a problem depends on whether the URL Google chose is one you are happy ranking; often it is, and the correct action is none, the same conclusion as in duplicate without user-selected canonical, where Google decides without your input.
When to act and when to leave it
Read the Google-selected canonical before touching anything. If Google chose a URL you are fine with, a www form, an HTTPS form, the parent of a filter page, then the content is indexed and served, just under a different address, and no traffic is lost. Acting on the status without reading the selection is how people spend a week changing tags that were never the problem.
Act when the selected canonical is genuinely wrong: a page you are retiring, a partner’s copy of your content, a thin variant. And then act on signals, not just the tag, because the tag already lost once. Align internal links and the sitemap on your preferred URL, redirect the loser where a redirect is honest, and give the pages enough distinct content that “duplicate” stops being true. On my own property the URL Inspection tool currently shows declared and selected canonicals agreeing on every page I inspect, which I verified again while writing this, and I keep it that way with boring consistency rather than clever tags: one URL form everywhere, sitemap included.
On Shopify this status has one overwhelmingly common cause, because every product on the platform ships with a second live URL under /collections/. If that is your platform, the Shopify version of this problem covers where those URLs come from and why the fix is internal links rather than a better canonical tag.
Sources
- Google Search Console Help, Page indexing report, status definitions and the four step diagnostic, checked 31 August 2026
- Google Search Central, How Google determines canonical URLs, on the hint and the competing signals
- Google Search Central, Consolidate duplicate URLs
- My own URL Inspection checks on shahidali.co, 31 August 2026